Article Isolation of SARS-CoV-2-related coronavirus from Malayan pangolins https://doi.org/10.1038/s41586-020-2313-x Kangpeng Xiao1,2,7, Junqiong Zhai3,7, Yaoyu Feng1,2, Niu Zhou3, Xu Zhang1,2, Jie-Jian Zou4, Na Li1,2, Yaqiong Guo1,2, Xiaobing Li1, Xuejuan Shen1, Zhipeng Zhang1, Fanfan Shu1,2, Received: 16 February 2020 Wanyi Huang1,2, Yu Li5, Ziding Zhang5, Rui-Ai Chen1,6, Ya-Jiang Wu3, Shi-Ming Peng3, Accepted: 28 April 2020 Mian Huang3, Wei-Jun Xie3, Qin-Hui Cai3, Fang-Hui Hou4, Wu Chen3 ✉, Lihua Xiao1,2 ✉ & Yongyi Shen1,2 ✉ Published online: 7 May 2020 Check for updates The current outbreak of coronavirus disease-2019 (COVID-19) poses unprecedented challenges to global health1. The new coronavirus responsible for this outbreak— severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)—shares high sequence identity to SARS-CoV and a bat coronavirus, RaTG132. Although bats may be the reservoir host for a variety of coronaviruses3,4, it remains unknown whether SARS-CoV-2 has additional host species. Here we show that a coronavirus, which we name pangolin-CoV, isolated from a Malayan pangolin has 100%, 98.6%, 97.8% and 90.7% amino acid identity with SARS-CoV-2 in the E, M, N and S proteins, respectively. In particular, the receptor-binding domain of the S protein of pangolin-CoV is almost identical to that of SARS-CoV-2, with one diference in a noncritical amino acid. Our comparative genomic analysis suggests that SARS-CoV-2 may have originated in the recombination of a virus similar to pangolin-CoV with one similar to RaTG13. Pangolin-CoV was detected in 17 out of the 25 Malayan pangolins that we analysed. Infected pangolins showed clinical signs and histological changes, and circulating antibodies against pangolin-CoV reacted with the S protein of SARS-CoV-2. The isolation of a coronavirus from pangolins that is closely related to SARS-CoV-2 suggests that these animals have the potential to act as an intermediate host of SARS-CoV-2. This newly identifed coronavirus from pangolins—the most-trafcked mammal in the illegal wildlife trade—could represent a future threat to public health if wildlife trade is not efectively controlled. As coronaviruses are common in mammals and birds5, we used the and crying. Furthermore, 14 of the 17 pangolins that tested positive for whole-genome sequence of SARS-CoV-2 (strain WHCV; GenBank acces- viral RNA died within one and half months of testing. Plasma samples sion number MN908947) in a Blast search of SARS-related coronavirus of four PCR-positive and four PCR-negative Malayan pangolins were sequences in available mammalian and avian viromic, metagenomic used in the detection of IgG and IgM antibodies against SARS-CoV-2 and transcriptomic data. We identified 34 closely related contigs in a using a double-antigen sandwich enzyme-linked immunosorbent assay set of viral metagenomes from pangolins (Extended Data Table 1), and (ELISA). One of the PCR-positive sample reacted strongly, showing an therefore focused our subsequent search on SARS-related coronavi- optical density at 450 nm (OD450) value of 2.17 (cut-off value = 0.11) ruses in pangolins. (Extended Data Table 2). The plasma remained positive at the dilution We obtained the lung tissues from 4 Chinese pangolins (Manis of 1:80, which suggests that the pangolin was naturally infected with a pentadactyla) and 25 Malayan pangolins (Manis javanica) from a wild- virus similar to SARS-CoV-2. The other three PCR-positive pangolins had life rescue centre during March–August 2019, and analysed them for no detectable antibodies against SARS-CoV-2. It is possible that these SARS-related coronaviruses using reverse-transcription polymerase pangolins died during the acute stage of disease, before the appear- chain reaction (RT–PCR) with primers that target a conservative region ance of antibodies. Histological examinations of tissues from four of betacoronaviruses. RNA from 17 of the 25 Malayan pangolins gener- betacoronavirus-positive Malayan pangolins revealed diffuse alveolar ated the expected PCR product, whereas RNA from the Chinese pango- damage of varying severity in the lung, compared with lung tissue from lins did not amplify. The virus-positive Malayan pangolins were all from a betacoronavirus-negative Malayan pangolin. In one case, alveoli were the first transport. These pangolins were brought into the rescue centre filled with desquamated epithelial cells and some macrophages with at the end of March, and gradually showed signs of respiratory disease, haemosiderin pigments, with considerably reduced alveolar space, including shortness of breath, emaciation, lack of appetite, inactivity leading to the consolidation of the lung. In other cases, similar changes 1Center for Emerging and Zoonotic Diseases, College of Veterinary Medicine, South China Agricultural University, Guangzhou, China. 2Guangdong Laboratory for Lingnan Modern Agriculture, Guangzhou, China. 3Guangzhou Zoo & Guangzhou Wildlife Research Center, Guangzhou, China. 4Guangdong Provincial Wildlife Rescue Center, Guangzhou, China. 5State Key Laboratory of Agrobiotechnology, College of Biological Sciences, China Agricultural University, Beijing, China. 6Zhaoqing Branch Center of Guangdong Laboratory for Lingnan Modern Agricultural Science and Technology, Zhaoqing, China. 7These authors contributed equally: Kangpeng Xiao, Junqiong Zhai. ✉e-mail: [email protected]; [email protected]; [email protected] 286 | Nature | Vol 583 | 9 July 2020 abe 3,798, 228, 669 and 1,260 bp, respectively, in length and the proteins they encode share 90.7%, 100%, 98.6% and 97.8% amino acid identity to the equivalent proteins of SARS-CoV-2 (Table 1). In a Simplot analysis of whole-genome sequences, we found that pangolin-CoV was highly similar to SARS-CoV-2 and RaTG13, with sequence identity between 80 and 98% (except for the S gene) (Fig. 2). cd Further comparative analysis of the S gene sequences suggests that there were recombination events among some of the SARS-related coronaviruses that we analysed. In the region of nucleotides 1–914, pangolin-CoV is more similar to the bat SARS-related coronaviruses ZXC21 and ZC45, whereas in the remaining part of the gene pangolin-CoV Fig. 1 | Pathological changes in the lungs of pangolins that are potentially is more similar to SARS-CoV-2 and RaTG13 (Fig. 2). In particular, the induced by pangolin-CoV. a–d, Histological changes in the lung tissues are receptor-binding domain (RBD) of the S protein of pangolin-CoV has compared between a virus-negative Malayan pangolin (a) and three only one amino acid difference with SARS-CoV-2. Overall, these data Malayan pangolins naturally infected with pangolin-CoV (b–d) (original indicate that SARS-CoV-2 might have originated from the recombina- magnification × 1,000). Proliferation and desquamation of alveolar epithelial tion of a virus similar to pangolin-CoV and a virus similar to RaTG13 cells and haemosiderin pigments are seen in tissues from all three infected (Fig. 2). To further support this conclusion, we assessed the evolu- pangolins and severe capillary congestion is seen in one of them (c). e, Viral tionary relationships among betacoronaviruses in the full genome, particles are seen in double-membrane vesicles in the transmission electron the RdRp and S genes, and in different regions of the S gene (Fig. 2c, microscopy image taken from Vero E6 cell culture inoculated with supernatant Extended Data Fig. 6). The topologies mostly showed the clustering of of homogenized lung tissue from one pangolin, with morphology indicative of coronavirus (inserts at the top right corner of e). Scale bar, 200 nm. pangolin-CoV with SARS-CoV-2 and RaTG13; SARS-CoV-2 and RaTG13 form a subclade within this cluster (Fig. 2c). However, pangolin-CoV and SARS-CoV-2 grouped together in the phylogenetic analysis of the were more focal (Fig. 1, Extended Data Fig. 1). The severe case also had RBD. Conflicts in cluster formation among phylogenetic analyses of exudate with red blood cells and necrotic cell debris in bronchioles and different regions of the genome serve as a strong indication of genetic bronchi. Focal mononuclear-cell infiltration was seen in the bronchi- recombination, as has previously been seen for SARS-CoV and Middle oles and bronchi in two of the cases, and haemorrhage was seen in the East respiratory syndrome coronavirus (MERS-CoV)6,7. bronchioles and small bronchi in one case (Extended Data Figs. 1–3). As the S proteins of both SARS-CoV and SARS-CoV-2 have previously Hyaline membrane and syncytia were not detected in the alveoli of the been shown to specifically recognize angiotensin-converting enzyme 2 four cases we examined. (ACE2) during the entry of host cells2,8, we conducted molecular bind- To isolate the virus, supernatant from homogenized lung tissue from ing simulations of the interaction of the S proteins of the four closely one dead Malayan pangolin was inoculated into Vero E6 cells. Obvious related SARS-related coronaviruses with ACE2 proteins from humans, cytopathogenic effects were observed in cells after a 72-h incubation. civets and pangolins. As expected, the RBD of SARS-CoV binds effi- Viral particles were detected by transmission electron microscopy: ciently to ACE2 from humans and civets in the molecular binding simu- most of these particles were inside double-membrane vesicles, with lation. In addition, this RBD appears to be capable of binding ACE2 of a few outside of them. They showed typical coronavirus morphology pangolins. By contrast, the S proteins of SARS-CoV-2 and pangolin-CoV (Fig. 1e). RT–PCR targeting the spike (S) and RdRp genes produced the can potentially recognize only the ACE2 of humans and pangolins expected PCR products: these PCR products had approximately 84.5% (Extended Data Fig. 7). and 92.2% nucleotide sequence identity, respectively, to the partial SARS-CoV-2 is one of three known zoonotic coronaviruses (the others S and RdRp genes of SARS-CoV-2.